5.Resources for human development

• - public sector prices, charges and costs
• - taxation and human development
• - foreign aid
• - decentralization and empowerment

One of the recurring themes of this study is that a decision to adopt a human development strategy does not imply an unusually large state or an unusually heavy burden of taxation. Human development is more concerned about spending priorities than about the total volume of expenditure. How the state spends its money is more important than how much money the state spends. Given that the total resources available to the government in most developing countries is a fifth to a quarter of total income, it should be possible to finance the state's contribution to human development by reallocating expenditure within exiting revenue ceilings without the necessity to raise additional revenues through taxation.

In many countries however the extent of reallocation will be quite considerable. Human development is an overall strategy of development and successful implementation of the strategy does not imply merely adding a few human development activities to the existing array of government spending programmes but, on the contrary, it implies a profound change in public policy and in the composition of state spending. Reallocations are likely to be necessary both within sectors (e.g., health) and between sectors (e.g., reduced spending on fertilizer subsidies and more spending on family planning services). These reallocations, once completed, often will result in a budget with a radically different shape from that found in the majority of developing countries.

There is of course no such thing as an ideal or optimal pattern of public expenditure that applies universally, independently of the circumstances of particular countries, and it would be fatuous to suggest otherwise. None the less, it might be useful to draw together the threads of the analysis in previous sections and present some orders of magnitude in order to visualize the broad contours of the strategy as a whole. This is presented in Table 2.

In constructing the "model budget" in Table 2 it is assumed that government revenue is equivalent to 25 per cent of gross national product. This is higher than the average in developing countries but below the ratios that can be found in countries such as Egypt (35.9 per cent), Zimbabwe (35.6 per cent), Tunisia (31.8 per cent), Chile (31.1 per cent), Malaysia (28.9 per cent) and Uruguay (28 per cent). Thus the assumption, while a little optimistic, is not unrealistic. Turning to the expenditure side, total spending has been dissaggregated into seven categories, four of which are directly concerned with human development (categories 1-4) and three of which are not (categories 5-7). The classification of expenditure used by the World Bank has been followed except for the guaranteed employment programme (category 4) which has been listed separately.

The allocations within the "model budget" are presented in two ways, namely, expenditure in each category expressed as a percentage of GNP and as a percentage of total government expenditure. The actual composition of central government expenditure in Indonesia is also presented so that comparisons can be made between the "model budget" and the revealed priorities of a large developing country following a more conventional strategy.

Note that in the "model budget" 66 per cent of government expenditure is allocated to human development activities (categories 1-4) whereas in Indonesia only 11.9 per cent of central government spending is channelled to these areas. Thus if Indonesia were to adopt a human development strategy and were to change its spending priorities to reflect new objectives, the share of expenditure allocated to education, health, social security and employment would have to rise by a multiple of five in order to approximate the priorities embodied in the "model budget". The changes contemplated are thus far from marginal.
 

Within the "model budget" equal priority is assigned to education, training, research and related activities (category 1) and to public health, nutrition and family planning (category 2), each accounting for about 20 per cent of total expenditure. Social security as a whole (category 3) is allocated 18 per cent of the budget, eight per cent assigned to food security, two per cent to economic security and eight per cent to all other programmes, including housing, sanitation and clean water. The guaranteed employment scheme (category 4) also is expected to account for about eight per cent of total government spending. Thus the three major structural reforms discussed in Section 4 which require sustained support--guaranteed employment, food security and economic security--absorb only 18 per cent of the budget. (It is assumed that asset redistribution--the other major reform--is a once-for-all measure which does not have recurrent budgetary implications. If it does, e.g., in the form of interest payments on government bonds used to compensate landowners, this would be funded from the allocations in category 7.)

It is envisaged in the "model budget" that 34 per cent of government expenditure (or 8.5 per cent of GNP) would be devoted to activities not directly connected to human development (categories 5-7). This is a much smaller proportion than one would expect to find in countries following a more conventional strategy of development. Indeed it is the sharp reduction in resources allocated to categories 5-7 that makes it possible to increase expenditure on human development. It is impossible to be specific about which programmes should be cut or eliminated, but it is clear in general that budgetary savings will largely arise from a combination of (i) a reduction in outlays on defence and related activities, (ii) the elimination of government subsidies to physical inputs, credit and services consumed by upper income groups, (iii) a sharp reduction and preferably the elimination of deficits in state economic enterprises and (iv) a contraction of expenditure on other government activities of relatively low priority. In addition, it is suggested below that human development is consistent with a more streamlined central government bureaucracy. Considerable savings may arise from doing away with the costly and unnecessary administrative machinery established to deliver goods and services to passive beneficiaries. These cuts and savings will not be easy, but they are essential because otherwise it will not be possible to reallocate government spending in favour of human development within a fixed revenue ceiling.

It has been assumed so far that government revenues are adequate and in the "model budget" specifically that total revenues are sufficient to finance an expenditure programme equivalent to 25 per cent of gross national product. In some countries however government revenue may not be sufficient to finance a human development strategy and it will consequently be necessary to raise additional funds.

Ignoring the possibility of external financing, there are several options from which to choose. First, the state could raise prices charged for the goods and services it produces for the public. This includes the prices charged by state economic enterprises as well as the charges levied by public utilities for power, water, drainage, telecommunications, etc., when these industries are in the public sector. This possibility was touched upon above in connection with eliminating the deficits of state enterprises, but the general issue is broader than this. In principle efficiency requires that the price of state produced goods and services cover long run marginal costs, including the opportunity cost of capital, and purchasers should expect to have to pay this price. In some cases this implies that state enterprises should behave in a more commercial, profit oriented way but without abusing their monopoly power. In other cases it may be better to transfer some of the state enterprises to the private sector by direct sale or auction. Provided the goods and services being supplied are consumed directly or indirectly by upper income groups, there is no conflict between allocative efficiency and distributive equity. However even when there is a conflict between the two objectives, policy makers should consider whether it would be possible to achieve the distributive objective in some other way, e.g., by increasing expenditure on human development activities oriented toward the poor.

Second, the government could consider imposing or increasing user charges to help recover some of the costs of supplying government services that often are provided free of charge. The main candidates here are health and education services. While favouring the introduction of user charges when this can be done without incurring heavy administrative costs or worsening the distribution of income, the efficacy of differential rate structures is doubtful because they tend to become overly elaborate and cumbersome. The objective should be to keep the structure of user fees simple, preferably with a single uniform rate and, where this is desired on grounds of equity, an exemption for members of a well defined group. User charges are perhaps most effective when the users are predominantly upper income persons and the costs of supplying the service are high. Examples include tertiary education and services provided by large urban hospitals.

Finally, the government can raise additional revenue by introducing new taxes or increasing the rates (up to a point) of old taxes. Here again, however, one should try to keep the tax system as simple as possible. There are numerous examples of governments in developing countries attempting to use the tax system to achieve multiple objectives, such as redistributing income, promoting investment, encouraging savings, accelerating growth and increasing efficiency. Most of these attempts have been unsuccessful and the country has been encumbered with a tax system that is complex, costly to administer and easy to evade. Experience therefore suggests that the wisest course, even for a human development strategy with a strong redistributive element, is to use the tax system primarily for what it does best, raising revenue. Taxation cannot by itself eliminate poverty or accelerate human development. These objectives are better addressed from the expenditure side of the budget. The role of the tax system is to raise revenue while making certain that the poor are not further impoverished and that human development is not impeded by government taxation.

One of the distinctive features of a human development strategy is that it is not intensive in the use of foreign aid or, indeed, of foreign capital in general. Human development is a strategy that mobilizes domestic resources and above all the energy, skills, talents and creativity of a nation's own people. Given that foreign aid and other forms of foreign capital are likely to be less readily available in future than in the past, human development is well suited for those countries which by choice or faute de mieux wish to become more self-reliant. Similarly, human development is not a strategy that is intensive in the use of foreign exchange: it does not rely heavily on imported capital and intermediate goods to sustain investment or production. The strategy is likely to be most successful where governments are committed to an open economy, biased neither towards export promotion nor import substitution.

Finally, a human development strategy does rely intensively on certain types of government services and support, above all on services that are in close contact with the grass roots: primary education, village health clinics, guaranteed employment on locally selected investment projects. This has implications for the administration of development and the empowerment of people. In the final analysis, if human beings are the end of development, if enhancing their capabilities and enlarging their choices is the ultimate objective, then empowerment must be central to what development is all about. Unfortunately the structure of governance in many developing countries--a structure often inherited from the colonial era--is antithetical to empowerment. The reason is that the state apparatus originally was designed to serve the purposes of a colonial administration, maintaining political control over a subject people, exploiting the territory's resources and extracting an economic surplus with the minimum amount of investment. Power was consequently highly centralized. This resulted in efficient administration in its own terms, but the structure of governance was inconsistent with political democracy and participatory development.

Governments in developing countries now have, or should have, new purposes but these new purposes have seldom been reflected in a restructured public administration. There is a strong case in many developing countries for bringing government closer to the people by devolving authority to lower levels of the administration, to provincial, district and local government. There is however a paradox here. Where once a highly centralized administration was an instrument of (colonial) tyranny, today the greater tyranny often is to be found at the local and provincial levels, where large landowners and their allies are able to exercise enormous power. Central government, in comparison, may be relatively benign and far sighted, more inclined to support human development in opposition to the entrenched forces that dominate the countryside. Thus decentralization is not an unmixed blessing or a goal in itself and it does not necessarily result in empowerment of the people. Decentralization is most consistent with the goals of human development when it is preceded by measures which redistribute wealth and income in favour of the poor and break the monopoly of power of local vested interests.

Decentralized government is not necessarily cheaper government. Indeed decentralization in support of human development could well result in a larger civil service and higher expenditure on general administration. The case for decentralization is not based on smaller government or lower costs but on improvements in the quality of government supplied services and on increased participation of the population in determining what is to be supplied. The intention, certainly, is to get a heavy handed central government "off the backs of the people", not in order to reduce costs while leaving the people to fend for themselves but in order to provide the essential public services needed to accelerate human development and increase the well being of the people.
 

Public sector prices, charges and costs

The public sector in many developing countries is quite large and supplies a wide range of goods and services from luxury international airline travel to nitrogenous fertilizer. The state enterprises that produce these goods and services often have to be subsidized because their receipts from sales fail to cover their total costs. Some of these subsidies, as discussed in Section 3, can be justified on efficiency grounds, because social costs and benefits differ from costs and benefits as measured by the market. Subsidies to state enterprises with important training, education or scientific research functions might fall under this category. In other cases the subsidies could perhaps be justified on grounds of equity, as a way of improving the distribution of income, but subsidies to state enterprises will almost always be a second best way to achieve distributive objectives. The remaining subsidies however, which account for most of the losses, can be justified neither on grounds of market failure nor on grounds of equity. State enterprises which incur such losses should be required to reduce their costs or raise prices or else they should be shut down or sold off.

Increasing the price of goods and services supplied by state enterprises can reduce claims on the exchequer to finance enterprise deficits and release funds for activities directly connected with human development. Similarly the government can impose user charges or fees for the services it provides. User charges however are likely to be appropriate only for certain types of services. In some cases, e.g. parks, beaches and other recreational facilities, nominal fees may be all that is required to prevent excessive use of the facilities and cover most of the day-to-day running costs. In the case of other natural resources which constitute part of the natural capital of a country, the issues are rather different, e.g. the optimal rate of exploiting fishery or timber resources, or the optimal rate of depleting petroleum, gas and other mineral deposits. Questions of conservation and the transformation of natural capital into physical and human capital come into play here. Some of the relevant principles were touched upon in section 1.

The case for user charges for government produced services, as opposed to charges for the use of government owned natural assets, is similar to the case for full pricing of products produced by state enterprises. The qualifications--externalities and equity--also are similar. Equity considerations suggest that user charges are highly appropriate for public services which benefit mainly middle and upper income groups. Tertiary education is a prime example and it is noteworthy that in South Korea user charges in higher education cover about 46 per cent of operating costs. The case for user charges for primary and secondary education is weak, although in China even the poor must make a token payment towards the cost of primary education and basic health services. Considerations of equity may tempt policy makers to introduce differential user charges for health and education based on the ability to pay. Such systems normally should be avoided: they are administratively costly and cumbersome and often fail to affect the target groups in the way intended.

Apart from increasing prices and imposing user charges, governments should attempt to lower costs of public sector activities. A human development strategy by its very nature--people-centred, participatory, democratic--is well suited to low cost methods of production and supply. The administration of primary schools and health clinics, for example, can be delegated to the local level, i.e., to the villages, small towns, communities and neighbourhoods where the facilities are located. Decentralization as such does not guarantee that costs will be lower--and in fact the advantages of economies of scale can sometimes be lost--but it does create a possibility for mobilizing locally available low cost resources, increasing the influence of intended beneficiaries, giving them a voice in what services are most needed and how they can best be supplied, and hence increasing the likelihood that the quality of the public services will improve. In this sense decentralization can be very cost-effective.

Participation is of course an end in itself; it is also a way of enhancing people's capabilities. Participation and its corollary, empowerment, enable people to exercise more control over their own lives. In allowing people to exercise control, governments help to unleash their energies, their commitment, their productivity and creativity. At the local level where people are most able to manage their own affairs, democracy clearly can help to enhance efficiency. In such a context, for example, teachers are more likely to become part of the community instead of a representative of an outside administration. Local people are more likely to be willing to supplement the salaries of teachers (possibly with payments in kind), to pay for some school supplies or to donate their own labour and materials to build and maintain a school. In such a context people cease to be objects of development and become the instruments of their own development, donating their time, energy and funds because they, their families and their friends will benefit directly.

In some cases the physical facilities, where they exist, can be used more intensively. For example, double shifts may be used in schools, thereby reducing the cost per student of physical capital. In other cases, less qualified workers, quickly trained, can be substituted for highly qualified workers whose training is lengthy and costly. The so-called barefoot doctors in China are an early example of how health workers with few formal medical qualifications can be used to extend the coverage and the effectiveness of an entire health care system. Teaching assistants or helpers can play the same role in primary schools. These are illustrations of how the costs of public services can be reduced.
 

Taxation and human development

Two additional points are critical: how can taxation be used to raise additional revenues when existing resources are inadequate and how should the tax system be reformed if it acts as an obstacle to human development. Governments usually have greater control over the allocation of expenditure than over the ways expenditure is financed. In particular governments in developing countries have found it difficult to tax people in accordance with their ability to pay. Upper income groups have been able to use their power and influence to get governments to subsidize activities of interest to themselves while at the same time ensuring that they are taxed relatively lightly. Thus both the income and expenditure sides of the state budget tend to favour the rich. The tax system often is regressive and upper income groups are able to escape paying their share by taking advantage of tax exemptions, income deductions and tax credits included in the revenue code as well as, of course, by outright tax evasion. It has turned out to be especially difficult to tax capital, in part because the valuation of capital assets such as land is difficult and in part because of the international mobility of financial capital.

Few tax systems in developing countries have succeeded in bringing about a significant redistribution of income by imposing a heavy tax burden on the rich. The richest five or ten per cent of the population almost always manage to avoid paying a large portion of the taxes for which they would appear to be liable. As a result, the post-tax distribution of income has a tendency to reflect the underlying distribution of wealth. Hence one reason for the emphasis on asset redistribution in Section 4. Some progressivity can be achieved however by exempting the lowest income groups from the payment of some taxes, notably income taxes. Given the administrative and political difficulties of using the tax system for redistributive purposes, a government pursuing a human development strategy probably would be well advised to concentrate on the distribution of the benefits from public expenditure and use the tax system to ensure that adequate revenues are available to finance planned expenditures. (i) import duties and export taxes

The base of a tax system can be almost anything--wealth, income, consumption, domestic or foreign trade--but many developing countries place heavy reliance on import duties and export taxes. The proportion may have fallen in the last decade or so, but many developing countries still obtain between a quarter and a third of their government revenue from taxes on international trade. Heavy taxation of internationally traded commodities--by increasing the cost of exporting and shielding import substituting industries from competition--tends to weaken the competitiveness of developing countries and makes their fiscal position vulnerable to fluctuations and instability originating abroad.

In Indonesia in the late 1970s and early 1980s, for example, two-thirds of tax revenue came from taxes on the export of oil and liquified natural gas. Had it not implemented a major tax reform before international oil prices fell sharply in the late 1980s, Indonesia would have encountered a fiscal crisis and been forced either to cut government expenditure sharply or tolerate huge fiscal deficits. 86 Similarly, two-thirds of Venezuela's public expenditure is financed by taxing exports of petroleum but unlike Indonesia, Venezuela did not introduce a tax reform and consequently the country ran into severe revenue problems when oil prices fell. A combination of low oil prices and depleted oil reserves does not augur well for the fiscal future of the country. 87

Other countries are highly dependent on revenues from import tariffs. The situation could be improved--revenue stability increased and international competitiveness enhanced--if governments were to shift the tax base toward domestic sources of revenue while maintaining low, uniform tariff rates on imports and exports. (ii) income taxes

In principle income taxes, and in particular taxes on personal incomes, should be attractive to governments which have adopted a human development strategy. In practice however personal income tax systems usually have not worked well in developing countries. Income taxes are attractive, first, because they can be made to be progressive and, second, because, if progressive, revenues rise faster than national income thereby imparting a degree of elasticity to the entire tax system. Progressive personal income tax systems, however, work best in countries where price inflation is low or moderate, the economy is fully monetized and accurate business records are kept, administration is honest and technically competent, literacy is universal, capital is immobile across international borders and tax compliance is voluntary. These conditions rarely are found in developing countries and as a result the revenue yield of income taxes has tended to be low. Income taxes are much more difficult to administer than indirect taxes on consumption and in those developing countries which have personal income taxes, seldom more than five per cent of the population is covered, primarily salaried persons working in the urban formal sector. 88 Income from property--retained profits, dividends, interest, rent--frequently is taxed at very low rates if at all. The tax system thus has a bias in favour of investments in physical capital (the income from which escapes taxation) and against human capital (the income from which, if sufficiently high, may be liable to income taxation). Erosion of the income tax base because of avoidance and evasion is common and severe, especially in the case of income from property. In Jamaica, for instance, it is estimated that because of inability to enforce tax legislation fully, actual revenues are only 54.4 per cent of potential revenues. If to the loss of revenue due to non-reporting of income one adds revenue lost because of loopholes in the tax code, then actual revenue falls to 38.2 per cent of potential revenue. 89 It is the self-employed who are in the best position to evade taxation and the incidence of evasion tends to increase with the level of income. As a result, a notionally progressive income tax system in fact is transformed into a regressive tax system.

Personal income taxes in developing countries thus tend to discriminate against income from human capital and to be inequitable. Income from physical and financial capital escapes most taxation. Income from labour, in contrast, is taxed relatively heavily particularly when taking into account the fact that the cost of accumulating human capital--the costs of education and training--are not allowed to be deducted from taxable income.

Taxes on corporate incomes and profits are easier to administer than taxes on personal incomes and in principle it should be possible to raise large revenues from a relatively small number of corporate taxpayers, domestic and foreign. However governments must beware of imposing very high average or marginal tax rates lest the after-tax rate of return on assets be pushed below what can be obtained abroad. This would almost certainly precipitate capital flight. The mobility of capital, in other words, imposes a limit on the extent to which taxation can be used for redistributive purposes. It is partly for this reason that the income tax system in most developing countries taxes physical and financial capital rather lightly compared to human capital. This introduces a bias against labour and human capital relative to other factors of production. (iii) taxes on wealth

Many of the problems that arise in taxing income also arise when taxing wealth. Since wealth and political power often go together, wealthy taxpayers in developing countries often succeed in shifting the tax burden from wealth to consumers or workers. The least difficult wealth taxes to collect are those on highly visible and immovable assets such as urban real estate and agricultural land. Taxes on urban property are sometimes said to be regressive, but this is unlikely to be true in most developing countries. Property taxes in fact are likely to be a more progressive way to cover the cost of providing urban services than progressively differentiated user charges. The reason for this is that the income elasticity of demand for housing (and other urban property) is greater than the elasticity of demand for the services themselves and hence spending on housing rises proportionately faster as incomes increase than spending on urban utilities and other services. The problem for the tax authorities is how to obtain and maintain accurate estimates of the value of urban property. Inevitably there is a conflict, on the one hand, between accuracy and fairness and, on the other, the administrative costs of valuation and tax collection. Once again, it is appropriate to keep it simple and use rough methods of valuation in order to keep administrative costs low.

Turning to the rural sector, the greatest difficulty policy makers confront in taxing agricultural land is undoubtedly political: the power of the landowning class. Where land is distributed highly unequally--a common phenomenon--the best solution, as argued in section 4, is a land reform. If land taxation were to accompany land redistribution, as we recommend, an equitable agricultural tax system could readily be established. Such a system, moreover, would be very broadly based and would provide revenue that could be collected at low cost.

Where land reform is impossible, it still is worthwhile introducing a land tax, provided of course that it is politically feasible. Even a very simple land tax can contribute significantly to government revenues while also being equitable. The key to land taxation is to base it on information that is readily available or that can be obtained at low cost. The revenue authorities need to know only four facts in order to design an effective system of agricultural land taxation: the area of the property, its location, its soil quality classification and the name of the owner. 90 In some countries the most difficult piece of information to obtain might appear to be soil quality, but recent developments in satellite imaging have made it much easier to estimate the potential fertility of the soil. In practice it may be more difficult to discover the name of the owner!

The easiest land tax to administer is one with a flat rate per unit of land with an exemption for holdings below a threshold size. If it is desired to have a progressive scale, then several rates per unit of land will have to be introduced with the rates rising with the total area of the property. This can quickly become complex to administer--particularly when landowners have several properties in different regions with different soil qualities--and we suggest that the fewer the number of tax rates, the better. Exemption of small holdings from taxation can go a long way toward making the system of land taxation moderately progressive. Exemption also ensures that the poor are not further impoverished by taxation. In Indonesia, for example, the property tax reforms of the 1980s included an exemption of US $1600 for all buildings, rural and urban, and the effect of this was to exclude from taxation virtually all houses in the rural areas and most houses occupied by low income families in the urban areas. 91 (iv) indirect taxes

During the last 10 to 15 years indirect taxes on consumption have become the principal means of raising revenue. Indeed in the 1980s indirect taxes accounted for more than half of government revenue in two-thirds of the developing countries. Value added tax in particular has been prominent in many of the most recent tax reforms and has been successful in raising revenue. The main concern about value added tax and other forms of indirect taxation is their effects on the post-tax distribution of income. Indirect taxation tends to be regressive.

The poor in developing countries, while usually not subject to income and wealth taxes, do nevertheless pay a significant amount of taxes. The urban poor, for instance, pay taxes equivalent to about 10 per cent of their income and much of this is accounted for by indirect taxes. A study of the Jamaican tax system revealed that the taxes which had the greatest impact on the poor were taxes levied on food and taxes on intermediate and capital goods used to produce food. Exemption of a few basic items such as cornmeal and condensed milk from the value added tax could go some way toward reducing the tax burden on low income households. 92 Thus it is possible to combine the efficiency of indirect taxes in raising revenue with measures to temper their regressive effects. Exemption from taxation of basic foods consumed largely by the poor would be consistent with human development and its goal of nutritional adequacy; it would also be desirable on grounds of distributive justice.

Excise taxes, such as those on beer, spirits and tobacco, pose interesting problems for a tax system concerned with poverty and human development. Excise taxes are very good for generating revenue because of the low price elasticity of demand of the goods taxed, but it is this low price elasticity which tends to result in a regressive impact of the taxes on the incomes of the poor, who buy large quantities of the taxed items. It sometimes is argued that high excise taxes on beer, spirits and tobacco reduce the ability of the poor to purchase basic necessities, including nutritious food. Against this is the argument that consumption of beer, spirits and tobacco produces large negative externalities (violence, death, accidents, illness) and the poor are no more entitled than the rich to inflict these heavy costs on the rest of society, including other members of the consumer's household. 93 Moreover, immoderate consumption of these items has the unfortunate effect of markedly depreciating the stock of human capital, reducing longevity, productivity and capabilities, and hence high excise taxes are not inconsistent with a human development strategy. (v) tax administration

There are many examples in developing countries of taxes with considerable revenue potential failing to deliver as expected simply because of weak tax administration. Part of the problem is technical: the taxes are too complex to administer or too difficult and costly to collect. Tax simplification often goes hand-in-hand with good tax administration. An obvious way to simplify a tax system is to eliminate many of the deductions, exemptions and credits that tend to become imbedded in the tax structure with the passage of time. If there are to be exemptions, the beneficiaries should be the poor rather than the rich. Yet unfortunately it commonly is the other way round.

Tax simplification can also be achieved by not incorporating non-revenue objectives into the tax system. Tax preferences given to favoured economic activities reduce total revenues while making the tax system more complex. They often conflict with one another or have effects which are incidental to or contradictory with the original objectives. Again, the rich invariably are in the best position to lobby for and benefit from preferential treatment. Governments which have adopted a human development strategy may be tempted to use the tax system to encourage human capital formation, e.g. by allowing deductions for the cost of training or formal education, but this objective probably can be better promoted by expenditure policies rather than tax policy. Tax gimmicks seldom are successful.

Ultimately the main reason tax administration often is weak and ineffective is political. Just as upper income groups are able to ensure that they receive a disproportionate share of government subsidies, so too they are able to ensure that they escape much of the burden of taxation. This pattern of subsidy and taxation occurs not by accident but by design. Experience has shown, however, that this inequity is easier to remedy by changing the composition of government expenditure than by changing the structure of the tax system. Governments wishing to promote human development are therefore well advised to concentrate their attention on the expenditure side of the budget, not only because human development objectives can be achieved this way with less difficulty but also because politically it may be unwise to fight a battle on two fronts simultaneously. The first priority should be to assemble a powerful political coalition to support a human development strategy on the basis of benefits arising from changes in government expenditure. Once this has been achieved, it should be possible to address the inequities imbedded in the tax system and by correcting these to bring about further improvements in the distribution of income and wealth. Governments which enjoy broad popular support are in the best position successfully to implement progressive tax reforms.
 

Foreign aid

During the last 45 years many developing countries have come to depend on external funds to finance their development programmes. Most of this external capital has consisted of some form of foreign aid--untied grants, bilateral tied grants and loans, surplus food products sold for local currency, multilateral loans provided at a subsidized interest rate, technical assistance--although in the 1970s loans from international commercial banks were significant and in a small number of countries direct private foreign investment has played a role. The results of foreign aid programmes, whether bilateral or multilateral, have in general been disappointing: they have done little to promote growth and even less to promote human development. 94

The rationale for foreign aid is that it is essential to compensate for an assumed shortage of domestic finance in developing countries. That is, foreign capital (in the form of foreign aid) supplements inadequate domestic savings and permits a country to have a higher rate of investment than would otherwise be possible. Greater investment, in turn, leads to faster growth. This chain of reasoning however is based on a false assumption, namely, that the chief obstacle in developing countries is a deficiency of physical capital and modern technology. The contrary view, the one advocated here, is that the principal obstacle to vigorous, sustainable growth in developing countries is the inadequacy of human development. Foreign aid has not in general been aimed at human development nor has it in practice encouraged human development indirectly. Inflows of foreign capital have tended to reproduce and reinforce the misallocations of domestic public expenditure that have been such a prominent feature of most government budgets.

Even if one believes, contrary to the accumulating evidence, that the engine of growth consists solely or primarily of investment in physical capital, it does not follow that developing countries are incapable of generating sufficient savings to finance that investment. Again, the evidence is clear that developing countries do have the capacity to attain high rates of savings and many have in fact done so. In 1990, for example, gross domestic savings in countries classified by the World Bank as low-income economies were 28 per cent of gross domestic product whereas the savings rate in the high-income economies (including the OECD countries) was much lower than this, viz., 22 per cent. 95 The problem in developing countries has not been an inability to generate savings for capital formation but to retain their capital domestically and prevent it from leaking abroad. There is a tendency in developing countries for capital to emigrate to more developed countries where it can obtain a higher rate of return. This tendency applies both to finance capital (liquid savings) as well as to human capital in the form of managerial, professional and technical personnel and of highly skilled and educated workers. 96

The meagre flows of foreign aid received by developing countries have sometimes had perverse effects, reducing the domestic rate of savings and encouraging capital flight by lowering the real rate of interest and temporarily appreciating the exchange rate. Hence foreign aid has contributed to the problem it was intended to correct. 97 Aid has been concentrated on a small number of countries. It has often been used to finance military expenditure, to enlarge the bureaucracy or to reduce the need for increased taxation. Where aid has resulted in increased investment in physical capital, it has typically financed large, capital intensive projects with relatively low rates of return. Where it has been directed at human development activities, concentration usually has been on the top of the expenditure pyramid, e.g., financing highly visible and expensive university facilities and urban hospitals. The social rates of return on these projects have been relatively low and the benefits have largely bypassed the poor.

Thus foreign aid is neither necessary in theory for a human development strategy nor has it been effective in practice. One can of course always hope that aid policies will be reformed and aid flows increased, but history provides little basis for such hopes. Human development, in contrast, permits a more self-reliant pattern of development. It is a strategy that mobilizes local resources, that concentrates on enlarging the latent capabilities of all people and that can be financed domestically without the need to depend on external subventions or charity. Indeed one of the attractive features of a human development strategy, for those governments which seek it, is that it liberates a country from dependence on foreign aid.
 

Decentralization and empowerment

It goes without saying that a human development strategy cannot be implemented if the revenues available to the government from domestic and foreign sources are inadequate. In addition, human development requires a public administration that is well adapted to the strategy and capable of implementing the tasks for which it is responsible. This essay has touched on several issues concerning the structure and role of state institutions in a human development strategy: the overall size of the state, the degree of decentralization and its relationship to participation and empowerment, the administration of tax reforms, etc. The following section pulls together the threads of the analysis and examines public administration in the round.

The central conclusion is that while the volume of work would be no higher, the intellectual demands placed on administrative capacities by human development will be greater than under more conventional strategies. The public administration will be asked to perform more complex tasks (e.g. to devise a calculus of benefits and costs of human development expenditure) and to relate to the public in a different way. The system of public administration consistent with a human development strategy is substantially different from that currently found in most developing countries. The strategy is not primarily concerned with social welfare intervention--the delivery of entitlements to passive beneficiaries--but with the empowerment of people and the liberation of their productive initiatives and creativity so that they can provide for themselves. Human development is about freedom and responsibility, liberation and self-help.

This is the context in which the role of the state should be seen. It is common these days for the state to be viewed as a source of inefficiency and a bureaucratic obstacle to development. This view is a reaction to an earlier one in which a highly centralized state was advocated with responsibilities for central planning and the direct management of major industries. The pendulum has now swung to the opposite extreme: wholesale privatization of state enterprises and a decentralized administrative structure are advocated. Supporters of human development have also argued in favour of decentralization, but their objective is fundamentally different from those who wish to construct a nightwatchman state with minimal powers and responsibilities.

As has been shown in this essay, human development requires an activist state, but an activist state which is able to enlist and mobilize people to promote their own development. The corollary of this is slightly paradoxical: insofar as the state pursues this objective successfully, it tends progressively to relieve itself of its traditional functions. The need for continuous intervention diminishes and with it the maze of bureaux required for monitoring, supervision and control that is such a prominent characteristic of state activity in most developing countries. In the long run, a human development strategy implies a gradual but decided reduction in the role of the state.

This will not occur because human development is best left to the unrestricted operation of market forces. Indeed market processes on their own are not capable of generating correct signals, of producing an efficient structure of incentives, if for no other reason than the web of positive externalities, complementaries and linkages associated with human development expenditures. No, the role of the state should gradually diminish because human development is best left to human beings themselves, as the direct participants in expanding their own capabilities. Some of the functions of the state gradually will be replaced by the institutions of civil society. In the short run, however, the objective of human development can be served either by a centralized administration or a highly decentralized public administration depending on the configuration of economic and political power in the country concerned.

Decentralization of the state should not be confused with empowerment of people. Depending on the circumstances, decentralization and empowerment can be in conflict, and in many cases they actually have been. It is perhaps useful to distinguish between administrative and political decentralization. 98 Most efforts to decentralize in developing countries have been administrative in character. That is, the primary purpose has been to increase the efficiency of government operations, not to enhance the political power of people. Political decentralization, in contrast, is not concerned primarily with increasing efficiency--improving the delivery mechanism of government services, removing bottlenecks and reducing delays, increasing the ability to recover costs--but with the devolution of power to the grass roots. Empowerment may result in greater efficiency, e.g., by mobilizing people in support of government initiatives and programmes, but that is not its purpose. Empowerment is an end in itself and should be judged in those terms and not in terms of its functional relationship to other worthy objectives such as cost-effectiveness or efficiency.

Decentralization cannot be divorced from the political context in which it occurs. If the people do not exercise democratic control over the central apparatus of the state, it is unlikely that decentralization of the state will be accompanied by increased political power of the people. If the political structure is undemocratic and authoritarian, administrative decentralization is likely either to maintain or even reinforce central authority. In the late colonial period, for instance, decentralization sometimes was used as a device to enlist the cooperation of local leaders and broaden the base of support for colonial rule. Similarly in South Africa, under pressure to become more democratic and less authoritarian, the apartheid state responded in the 1980s by introducing a programme of regional and urban decentralization. While ostensibly democratic, the real intent of the administrative reforms was to increase the power of the centre over local government, depoliticize the black population and produce a new group of political collaborators. 99 Those who opposed apartheid and aspired to a democratic society resisted administrative decentralization and concentrated their efforts on gaining control of the central state apparatus itself.

Consider a less extreme case, that of Nepal. Decentralization within this country of 18.9 million persons frequently has been recommended both by outside consultants and officials within the government. One reason for the recommendation is the many inefficiencies that arise from attempting to administer a very backward economy through a highly centralized political authority. Yet the experience of Nepal illustrates why administrative decentralization is doomed to failure in the absence of a devolution of political power.

Nepal is highly stratified, economically, socially and politically. About 90 per cent of the labour force is engaged in agriculture and lives in rural areas. The ownership of land however is highly concentrated and the top three per cent of landowners possess nearly 40 per cent of the land. As a result, the distribution of income in Nepal is one of the most unequal in Asia. Decentralization in Nepal has historically been supported by two groups: populist opponents of the regime and modernist administrators in government. The populists supported decentralization because they believed it would result in a reduction in the power of the central bureaucracy in the capital and encourage grass roots development. The modernizing administrators supported decentralization as a way to respond to pressure for more local participation without having to legalize opposition political parties. That is, administrative decentralization was viewed as a tactic to defuse demands for more radical social and political change.

Not surprisingly, decentralization in Nepal has done little to encourage popular democracy. Villages and districts have little voice and exercise little influence over national policies. Such grass roots participation as exists is controlled and manipulated by the rural elite and local government officials. At the village and ward levels the political system is dominated by rich farmers. 100 If the distribution of wealth and income at the local level are highly unequal and political authority is highly concentrated, as in Nepal, administrative decentralization is unlikely to result in political decentralization and the empowerment of the population. Conservative landed elites usually dominate local politics in agrarian societies. The outcome of a policy of decentralization in such societies normally is doubly disadvantageous, namely, very little actual administrative decentralization plus the capture by large landowners of whatever political decentralization occurs. Compared to this, a centralized administration may be preferable.

In the final analysis human development cannot be separated from participatory development. The goal of human development is to help people realize their own potential, to develop their intellectual, technical and organizational capabilities. Thus human development inescapably is development by the people if it is to be development for the people. 101 In order to fulfil their potential people must participate actively in constructing their own autonomous, democratic organizations, including of course their political organizations. Political empowerment is an integral aspect of participatory development. Those from outside the community--whether representatives of government or non-governmental organizations--who help people construct their grass roots institutions can be most effective when they function chiefly as facilitators, catalysts or animators. An activist state can encourage or animate human development but it cannot engineer it. That must be done by the people themselves and experience suggests that direct democracy at the local level is the best way to foster it. The role of catalyst or animator, depending on the configuration of political power, can be played either by the institutions of local government or by a centralized state operating through mass mobilization campaigns, i.e., mass literacy campaigns, mass immunization campaigns, mass rural construction campaigns and the like. It would be wrong to assume that local bureaucracies are more sympathetic to grass roots democracy than centralized bureaucracies. Experience suggests however that a programme of decentralization is most successful when local government works in concert with active institutions of civil society, such as community organizations, trade unions and peasant associations.

The connections at the local level between direct democracy and human development are fairly clear. Difficulties arise when large scale projects are considered and the geographical area encompasses a number of communities or villages. The administration of secondary schools or regional hospitals are examples of what we have in mind. Once one moves beyond the local level more complex institutions become necessary and direct democracy inevitably has to give way to representative democracy. That is, once one moves beyond the local level it becomes necessary to delegate authority to leaders who are entrusted to safeguard and advance the interests of a fairly large group. Decision making inevitably becomes more centralized but this need not be in conflict with direct democracy at the community level. Indeed, the centralization necessary for the minimal coordination of governmental activities should be used to facilitate local democracy.

While the historical evolution of national institutions and the concrete political and economic realities of each country will differ, the model which would appear to be most supportive of a human development strategy is one which promotes democracy "from the bottom up" rather than "from the top down". This is because a human development strategy places special emphasis on small, widely dispersed, labour intensive investment activities such as primary schools, community health clinics, micro-irrigation schemes and local sanitation systems. Since it is at this grass roots level that direct democracy is most likely to flourish, it makes sense on both economic and political grounds for a state committed to human development to direct its efforts toward strengthening the basic organs of self-government in the rural villages and urban neighbourhoods. The local institutions of civil society in conjunction with local government organizations can function as "schools of democracy" in which people acquire experience in taking control of their own lives, individually and collectively. Once the foundation is secure, other floors in the edifice of democracy can be constructed at ever higher levels.

The opposite approach is to try to promote democracy "from the top down". This approach has become widespread in developing countries in recent years and the experience of Bangladesh is illustrative of the consequences. The central government has decentralized some decision making and revenue generating functions to intermediate regional levels, namely, to districts and sub-districts, but it has not increased the responsibilities and authority of the basic unit of local government, the Union Council, although it has a long history. Local government, in other words, continues to be of marginal significance even after decentralization. Grass roots organizations have in the past filled part of the void and have taken responsibility for the maintenance of local roads, small irrigation systems, primary schools, local water supply and village health and nutrition services. Rather than build on these embryonic institutions and use them to promote human development, the central government has ignored and neglected them and in many instances allowed them to die. Despite this unfortunate history new NGOs have emerged and in some cases have achieved considerable success. The best known examples are the Bangladesh Rural Advancement Committee, the Grameen Bank and Proshika. Evidently the tradition of local, self-help organizations has not disappeared completely and with support from government local institutions could perhaps be revived and used to accelerate human development. 102

The central government in Bangladesh has tended to distance itself from the non-governmental organizations whereas it should instead regard the efforts of NGOs as complementary to its own. With a little encouragement it is likely that NGOs could provide even more assistance to grass roots organizations than they do at present. As it is, 55 per cent of the total budget of NGOs in Bangladesh is allocated to human development activities. A vigorous civil society consisting of independent organizations such as NGOs, trade unions, farmers associations, women's associations and youth organizations can contribute much to a broadly based programme in support of human development. In Bangladesh however the success of the NGOs in advancing human development merely underlines the government's own lack of initiative. As far as human development is concerned, decentralization in Bangladesh has been a red herring.

It is often claimed that administrative decentralization results in a more efficient delivery of public services than is possible under a centralized system. This claim however should be regarded with scepticism. Decentralization in practice often consists of little more than an enlargement and extension of the government's bureaucracy, a multiplication of posts and offices at lower levels of administration. Problems of coordination and communication often become worse: delays increase, costs mount. Decentralization is associated not with streamlined administration but with a proliferation of bureaucracy. There is thus a danger that decentralization will promote neither democracy nor efficiency. Empowerment, on the other hand, is not only democratic, it may also be efficient. It calls for a streamlined central administration which devolves authority as much as possible to the local level. By enlisting the active participation of the people instead of relying on a cumbersome bureaucracy to "deliver" services to beneficiaries who have little voice in what is delivered and how it is delivered, empowerment may actually be more cost-effective than the alternatives. The experience of West Bengal, India suggests that this may have happened there. 103 Where this does occur, empowerment, human development and economic efficiency are inextricably intertwined. They are the principal components of a strategy that puts people first.

NOTES

86. Malcolm Gillis, "Comprehensive Tax Reform: The Indonesian Experience, 1981-1988," in Malcolm Gillis, ed., Tax Reform in Developing Countries, Duke University Press, 1989.

87. Charles E. McLure, Jr., "Income Tax Reform in Colombia and Venezuela: A Comparative History," World Development, Vol. 20, No. 3, March 1992.

88. Malcolm Gillis, "Tax Reform: Lessons from Postwar Experience in Developing Nations," in Malcolm Gillis, ed., op. cit.

89. James Alm, Roy Bahl and Matthew N. Murray, "Tax Base Erosion in Developing Countries," Economic Development and Cultural Change, Vol. 39, No. 4, July 1991.

90. David Newberry, "Agricultural Taxation: The Main Issues," in David Newberry and Nicholas Stern, eds., The Theory of Taxation for Developing Countries, New York: Oxford University Press, 1987.

91. Malcolm Gillis, "Comprehensive Tax Reform: The Indonesia Experience, 1981-1988," loc. cit.

92. Richard M. Bird and Barbara Diane Miller, "The Incidence of Indirect Taxes on Low-Income Households in Jamaica," Economic Development and Cultural Change, Vol. 37, No. 2, January 1989.

93. Richard M. Bird, Tax Policy and Economic Development, Baltimore: Johns Hopkins University Press, 1992.

94. UNDP, Human Development Report 1992, New York: Oxford University Press, 1992.

95. World Bank, World Development Report 1992, New York: Oxford University Press, 1992, Table 9, pp. 234-5.

96. This argument is developed in Keith Griffin and Azizur Rahman Khan, Globalization and the Developing World, Geneva: United Nations Research Institute for Social Development, 1992, Ch. 4.

97. Keith Griffin, "Foreign Aid After the Cold War," Development and Change, Vol. 22, No. 4, October 1991.

98. See the valuable discussion in Joel Samoff, "Decentralization: The Politics of Interventionism," Development and Change, Vol. 21, No. 3, July 1990.

99. Ibid.

100. Henry Bienen, Devesh Kapur, James Parks and Jeffrey Riedinger, "Decentralization in Nepal," World Development, Vol. 18, No. 1, January 1990.

101. See Dharam Ghai, "Participatory Development: Some Perspectives from Grass-Roots Experiences," loc. cit.

102. UNDP, Human Development in Bangladesh, loc. ci.